Enantiomeric resolution of albuterol sulfate by preferential crystallization

Article abstract of DOI:10.1016/j.tetasy.2007.05.010

Albuterol is a β₂-adrenoceptor agonist prescribed for the treatment of bronchial asthma; it exists as a racemate and its bronchodilator activity resides in the (R)-isomer or levalbuterol. The aim of this study was to determine a methodology that would separate the enantiomers of albuterol by preferential crystallization after a conglomerate is identified within its derivatives. We found that albuterol sulfate behaves as a conglomerate showing the characteristic α_x-value = 2 (mole fraction solubility ratio of racemate vs enantiomer), the V-shaped ternary phase diagram and the preferential crystallization by seeding with the pure enantiomer. On the basis of these characteristics, we separated the enantiomers by entrainment, and crystallizing out a saturated methanolic solution of albuterol sulfate at 15 °C.

Full text of DOI:10.1016/j.tetasy.2007.05.010

Tetrahedron: Asymmetry 18 (2007) 1170–1175
Enantiomeric resolution of albuterol sulfate
by preferential crystallization
Sara M. Palaciosa,* and Marcela A. Palaciob
aFacultad de Ciencias Quı´micas, Universidad Cato´lica de Co´rdoba, Cno. a Alta Gracia Km 10, 5000 Co´rdoba, Argentina
bAgencia Co´rdoba Ciencia SE, Unidad CEPROCOR, 5164, Santa Marı´a de Punilla, Co´rdoba, Argentina
Received 22 March 2007; accepted 15 May 2007
Abstract—Albuterol is a b2-adrenoceptor agonist prescribed for the treatment of bronchial asthma; it exists as a racemate and its bron-
chodilator activity resides in the (R)-isomer or levalbuterol. The aim of this study was to determine a methodology that would separate
the enantiomers of albuterol by preferential crystallization after a conglomerate is identified within its derivatives. We found that albu-
terol sulfate behaves as a conglomerate showing the characteristic ax-value = 2 (mole fraction solubility ratio of racemate vs enantiomer),
the V-shaped ternary phase diagram and the preferential crystallization by seeding with the pure enantiomer. On the basis of these char-
acteristics, we separated the enantiomers by entrainment, and crystallizing out a saturated methanolic solution of albuterol sulfate at
15 ꢀC.
ꢁ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
(c) pseudoracemate.6 Only 5–10% of racemates belong
to the conglomerate forming group, which is the most
favorable group for achieving a certain enantiomeric
enrichment by fractional crystallization;7 the search for
or the discovery of new conglomerates among the chiral
drugs, synthons, or their precursors has proven to be quite
fruitful.8
Chirality is important in the context of biological activity
because at a molecular level, asymmetry dominates the bio-
logical processes.1,2 As the essential components of life
such as proteins, carbohydrates, or DNA are constructed
from enantiomerically pure building blocks, great differ-
ences are observed in the activity of enantiomeric drugs,
and this results in different bioactivity, bioavailability,
pharmacokinetics, and pharmacodynamics.2 Conse-
quently, regulatory authorities increasingly demand that
chiral drugs should be administered in an enantiomerically
pure form.3 This has provoked tremendous interest within
the pharmaceutical industry in obtaining pure enantiomers,
either through enantioselective synthesis or the subsequent
enantioseparation of racemic mixtures.4
Albuterol Al, also known as ( )-2-tert-butylamino-1-(4-
hydroxy-3-hydroxy-methyl) phenyl ethanol, is a b2-adreno-
ceptor agonist prescribed for the treatment of bronchial
asthma.9 Al is a racemate and its bronchodilator activity
resides in the (R)-isomer or levalbuterol.10,11
In 1997, under the chiral switch strategy,12 levalbuterol
hydrochloride and levalbuterol sulfate (RAS, Fig. 1) were
approved by the FDA.13 Since 1999 levalbuterol hydro-
chloride has been marketed as a nebulizer solution and in
Classical crystallization techniques continue to be one of
the most commonly used methods in industrial resolutions
of racemic mixtures.5 In order to achieve an efficient sepa-
ration of enantiomers by crystallization, the nature of the
racemate must be known. Three types of racemic mixtures
have been identified on the basis of their solid–liquid phase
diagrams: (a) conglomerate, (b) racemic compounds, and
OH
H
N
H2SO4
HO
OH
2
*
Corresponding author. Tel.: +54 351 4938000x611; fax: +54 351
4938061; e-mail: sarapalacios@campus1.uccor.edu.ar
Figure 1. (R)-Albuterol sulfate (RAS).
0957-4166/$ - see front matter ꢁ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2007.05.010

S. M. Palacios, M. A. Palacio / Tetrahedron: Asymmetry 18 (2007) 1170–1175
1171
metered dose inhalers. Efforts to make a safer albuterol sul-
fate (AS) drug have prompted the need for new methods to
obtain RAS. Herein, we report that albuterol sulfate crys-
tallizes as a conglomerate and its resolution by preferential
crystallization.
2. Results and discussion
In order to find a conglomerate within the Al derivatives,
we made a series of determinations and experiments to sup-
port the presence of a conglomerate. First, we determined
the eutectic composition of albuterol (Al), albuterol HCl
(AHC), and albuterol sulfate (AS). The eutectic composi-
tion of Al was R/S = 75/35, indicating that it is a racemic
compound. AHC did not crystallize or separate from the
mother liquor as a gel material and so we were unable to
determine its eutectic composition. AS, which has good
crystal behavior, showed a eutectic point equal to R/
S = 50/50. This result indicated that we were possibly in
the presence of a conglomerate; a series of experiments
were therefore carried out.
Figure 2. TPD of AS.
2.1. Preferential crystallization by entrainment of
(R)-albuterol sulfate
Before performing the crystallization experiments, the opti-
mum values for the different crystallization variables were
evaluated and determined at a concentration of 16–18 g/l,
stirring at 450–500 rpm, and a crystallization temperature
of 15 ꢀC. Seeding was also evaluated and was considered
unnecessary, since an enriched solution of AS (17 g/l, R/
S = 55/45) seeded with 0.5% of RAS gave a 24% yield of
RAS with an enantiomeric excess of R/S = 76.5/23.5, while
the same solution without seeding, yielded 29% of RAS and
R/S = 80/20.
Second, the solubility, determined in methanol at 25 ꢀC, of
AS and RAS was 8.77 and 4.00 mg/ml, respectively. As
this was consistent with the presence of a conglomerate,7
we were able to determine a 2.22 ratio ax between the
solubility of AS and RAS (both expressed as mole frac-
tions) while similar ax-values were found at 30 and 40 ꢀC
(Table 1).
Table 1. Experimental solubility (molar fraction) of RAS, AS and ax ratio
Temperature (ꢀC)
RAS (·105)
AS (·105)
ax Ratio
In the first cycle of successive crystallizations, a highly
supersaturated solution of AS (4.04 g in 230 ml of metha-
nol), enriched with an excess of RAS (0.46 g, R/S = 55.1/
44.9) was stirred for 1 h at 15 ꢀC. Afterwards, we isolated
1.11 g of RAS (R/S = 82.7/17.3) (Table 2). According to
Amaird’s method,7 the optimum yield of enantiomeric
crystals in each cycle will be double the amount of the
excess RAS mass. In our case, we found 0.72 g of RAS
(according to the optical purity) was obtained, which is
nearly 79% of the theoretical yield (0.92 g). Then 1.00 g
of AS was added to the resulting (S)-enriched mother
liquor to restore the initial supersaturation and the proce-
dure was repeated again to obtain the (S)-enantiomer
(1.34 g, R/S = 22.2/77.8). After four cycles for each enan-
25
30
40
2.80
33.08
52.41
6.23
73.58
108.58
2.22
2.22
2.07
As a third confirmation, a typical preferential crystalliza-
tion experiment was performed. Thus, 0.236 g of AS was
dissolved in 20 ml of MeOH at reflux, then allowed to
cool down, and seeded (1% w/w) with pure RAS. After
12 h, we decanted the mother liquors, isolating a precipi-
tate with an enantiomeric composition of R/S = 95/5 in a
5% yield.
Prompted by the successful resolution of AS through pref-
erential crystallization, we determined the ternary phase
diagram (TPD) of AS instead of the binary phase diagram
(BPD), because AS decomposes before it melts, thus pre-
venting the determination of melting points.14
Table 2. Resolution of AS by entrainment
Cycle no.
Mass added
(g)
Recovery of resolved AS in g
(R/S percentage)
The solubility of AS, RAS, and mixtures of both com-
pounds determined in methanol at 25, 30, and 40 ꢀC are
graphically represented in Figure 2. The V-shaped curve
of the TPD confirmed the finding of a conglomerate.6,7
AS
RAS
0.46
RAS crystals
SAS crystals
1
4.04
1
1
1.5
0.70
0.80
1
0.80
10.84
1.11 (82.7/17.3)
1.34 (22.2/77.8)
0.74 (19.6/80.4)
0.99 (31.6/68.4)
2
1.5 (60.3/39.7)
0.80 (72.9/27.1)
0.80 (71.7/28.3)
4.21 (70.8/29.2)
The eutectic composition of AS, the a-value, the V-shaped
TPD, and the preferential crystallization, indicated that AS
shows conglomerate characteristics and that separation by
entrainment is a suitable methodology for separating its
enantiomers.
3
4
0.99 (32.2/67.8)
4.06 (26.5/73.5)
Total
0.46

1172
S. M. Palacios, M. A. Palacio / Tetrahedron: Asymmetry 18 (2007) 1170–1175
tiomer, we obtained 4.21 g with R/S ꢀ 70/30 and 4.06 g R/
S ꢀ 30/70 (Table 2). Finally, this RAS quantity was twice
recrystallized from methanol, yielding 1.65 g of RAS (R/
S = 99.8/0.2).
as we know, this situation in which the characteristic solu-
bility of the conglomerate racemate was not confirmed by
crystal spectroscopy is unusual. Thus, we are attempting
to follow up this experiment with a study of the single crys-
tal structure in order to elucidate this discrepancy.
The RAS crystals were characterized spectroscopically, as
shown in Figures 3–5. The analysis of the liquid NMR
spectra (Fig. 3) and DSC thermogram14 indicated that
the RAS crystals did not correspond to solvates or
hydrates; therefore, the conglomerate crystallization corre-
sponded to pure RAS.
3. Conclusion
Herein, we have reported a continuous cycle process for
converting racemic AS into the chiral anti-asthmatic drug
RAS, on the basis of the discovery of its conglomerate
characteristics. The resolution of AS by preferential crys-
tallization according to the entrainment procedure was
tried and successfully developed by preventing the seeding
of the saturated solution before each crystallization. A
clear advantage of this methodology is to produce RAS
from the widely available racemic AS without chiral auxil-
iaries, except for the minimum initial RAS enrichment.
The XRPD pattern of RAS obtained in the preferential
crystallization cycles, showed characteristic peaks at 8.7;
9.6 and 15.2 (2h 0.2ꢀ 2h) (Fig. 4c), corresponding to the
RAS polymorphic Form II as reported by Palacio et al.14
A characteristic of the conglomerates is that the XRPD
pattern of the pure enantiomer is identical to that of the
racemate.6 However in this case, we observed different pat-
terns in the AS (racemate) (Fig. 4a) and the RAS (enantio-
mer). RAS presented two polymorphic forms and neither
the RAS Form II nor RAS Form I (Fig. 4b) were similar
to the racemate pattern.15 In spite of the fact that there
has been no report of the occurrence of polymorphism in
AS, we looked for other crystal modifications for the
racemate but found none. Clear differences were observed
between the IR spectra of AS,16 RAS Form II and RAS
Form I14 (Fig. 5a–c), confirming the difference in their crys-
tal structure. In light of this evidence, we concluded that
RAS Form II, RAS Form I and AS crystallize in different
crystal forms.
4. Experimental
4.1. Materials
RAS was prepared as described in a previous report17 and
the purity of this compound was 99.57% assayed against
Albuterol Sulfate USP standard, with an enantiomeric ex-
cess of 99.8% by HPLC.18 AS was provided by Laboratorio
Pablo Cassara, S.R.L. (Buenos Aires, Argentina). All sol-
vents used in this study were of HPLC grade (Fisher Scien-
tific, New Jersey).
These findings (XRPD and FT-IR spectra for AS being dif-
ferent from those for RAS) indicate that AS does not crys-
tallize as the classical mixture of enantiomeric crystals
expected for a conglomerate.7 However, its solubility
behavior in methanol (eutectic at R/S = 50/50, ax-value
ꢀ2, preferential crystallization, V-shape TPD) allows us
to separate it through preferential crystallization. As far
4.2. Methods
4.2.1. Determination of eutectic composition. In a 25 ml
erlenmeyer flask, 200 mg of solid material (R/S = 60/40
and R/S = 70/30) was left to be solubilized in 10 ml of
methanol at 25 ꢀC. After 24 h, an aliquot of the superna-
/2
7.5
7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
(ppm)
Figure 3. NMR spectra of RAS and AS.

S. M. Palacios, M. A. Palacio / Tetrahedron: Asymmetry 18 (2007) 1170–1175
1173
0
10
20
30
40
50
2 Tetha
0
10
20
30
40
50
2 Tetha
0
10
20
30
40
50
2 Tetha
Figure 4. XRPD patterns of (a) AS, (b) RAS Form I and (c) Form II.
tant was filtered in a HPLC filter. The composition of the
allowed to solubilize in 10 ml of dried methanol at 25,
30, and 40 ꢀC. The erlenmeyers were kept in a thermostatic
bath regulated at the desired temperature 1 ꢀC. After
24 h, 1 ml of the supernatant was filtered in an HPLC filter.
Then, the amount dissolved and the R/S composition of
solid material were quantified by chiral HPLC.18
dissolved R/S was then analyzed by chiral HPLC.18
4.2.2. Solubility determination of albuterol sulfate and its
enantiomer in methanol. In a 50 ml erlenmeyer flask, an
excess amount of the corresponding solid material was

1174
S. M. Palacios, M. A. Palacio / Tetrahedron: Asymmetry 18 (2007) 1170–1175
2.0
1.0
0.0
2.0
4000
3000
2000
1500
Wavenumber cm-1
1000
500.0
1.0
0.0
4000
3000
2000
1500
Wavenumber cm-1
1000
500.0
2.0
1.0
0.0
4000
3000
2000
1500
1000
500.0
Wavenumber cm-1
Figure 5. FT-IR spectra of (a) AS, (b) RAS Form I and (c) Form II.
4.2.3. Preferential crystallization by entrainment of (R)-
albuterol sulfate. A reflux solution of AS (4.04 g) and
RAS (0.46 g) in methanol as solvent (230 ml) was cooled
down to 15 ꢀC to give a supersaturated solution, which
was stirred at 500 rpm for 1 h. A precipitate then appeared
and was collected by filtration, washed with cool methanol,
and dried under vacuum, to give 1.11 g of RAS (R/
S = 82.7/17.3). Next, 1 g of AS was added to the filtrate
and dissolved at 65 ꢀC. The solution was similarly cooled
to 15 ꢀC and stirred gently for 1 h, to yield 1.34 g of SAS
(R/S = 22.2/77.8). This cycle was repeated four times, with
the addition of 0.8–1 g of racemate after each crystalliza-
tion in order to restore the concentration of the solution.
After several stages of the preferential crystallization, the
crystals with the same specific rotation were combined.
Recrystallization of RAS (4.21 g, R/S = 71/29) from meth-
anol gave 49% of highly pure RAS (2.06 g, R/S = 93/7). In a
similar manner, SAS (4.06 g, R/S = 26/74) was crystallized
from methanol giving 2.07 g (51%) of SAS (R/S = 8/92). In

S. M. Palacios, M. A. Palacio / Tetrahedron: Asymmetry 18 (2007) 1170–1175
1175
a third process of crystallization from methanol, 2.06 g of
RAS (R/S = 93/7) gave 1.65 g of RAS (R/S = 99.8/0.2),
representing an 80% yield. The overall yield of RAS was
15%, calculated on the basis of 50% theoretical yield.
heated in crimped aluminum pans from 30 to 300 ꢀC at a
rate of 10 ꢀC minÀ1 under static air.
Acknowledgments
4.2.4. HPLC analysis. A Waters 2690 HPLC system with
quaternary pump and autosampler and a Waters 996 pho-
todiode array detector were used. For data acquisition,
MILLENNIUM 3.20 software was used. All separations were
achieved using a 25 · 4.6 mm Chirobiotic T column
(amphoteric glycopeptide Teicoplanin bonded to a 5 lm
silica gel) from ASTEC (Whippany, New Jersey).
´
The authors would like to thank Universidad Catolica de
´
´
Cordoba and Agencia Cordoba Ciencia for their financial
support.
References
All samples and standard solutions were chromatographed
at ambient temperature (22 2 ꢀC) using an acetonitrile/
methanol/acetic acid/triethylamine mixture (60/40/0.3/0.2
v/v/v/v) as the mobile phase (flow rate of 1.5 ml minÀ1),
with detection at 276 nm and an injection volume of 10 ll.18
1. For examples, see the following: (a) Bentley, R. Chem. Rev.
2006, 106, 4099–4112; (b) Rouhi, A. M. Chem. Eng. News
2003, 81, 45–55; (c) Eliel, E.; Wilen, S. Stereochemistry of
Organic Compounds; John Wiley & Sons: New York, 1994; pp
201–214.
2. Sheldon, R. Chirotechnology, Industrial Synthesis of Optically
Active Compounds; Marcel Dekker: New York, 1993; pp 39–
72.
3. Lorenz, H.; Perlberg, A.; Sapoundjiev, D.; Elsner, M. P.;
Seidel-Morgentern, A. Chem. Eng. Proc. 2006, 45, 863–873.
4. Stinson, S. C. Chem. Eng. News 1995, 44.
5. Wood, W. M. L. Crystal Science Techniques in the Manu-
facture of Chiral Compounds. In Chirality in Industry II;
Collins, A. N., Sheldrake, G. N., Crosby, J., Eds.; John
Wiley: New York, 1997; pp 119–156.
4.2.5. Solution-state NMR. NMR spectra were obtained
in a Bruker NMR with a Bruker AC 200 console (Bruker,
Germany). The spectra were processed with WinNMR
software (Bruker, Germany). The samples were prepared
by dissolving 5 mg of each form in 0.5 ml of DMSO-d6 with
0.03% of tetramethylsilane (TMS, Sigma–Aldrich Chemi-
cal Co., Wisconsin) used as reference for dH = 0.
4.2.6. Infrared spectroscopy. Fourier transform infrared
(FTIR) spectra were acquired on a Shimadzu spectrometer
(Shimadzu, Kyoto, Japan). Spectra over a range of 4000–
500 cmÀ1 with a resolution of 2 cmÀ1 (50 scans) were re-
corded using KBr pellets. For diffuse reflectance analysis,
samples weighing approximately 2 mg were mixed with
200 mg KBr using an agate mortar and a pestle, and were
placed in sample cups for fast sampling.
6. Jacques, J.; Collet, A.; Wilen, S. In Enantiomers, Racemates
and Resolution; Krieger: Malabar, 1994; pp 32–57.
7. Jacques, J.; Collet, A.; Brienne, M. J. Chem. Rev. 1980, 80,
215–230.
8. Perez-Garcia, L.; Amabilino, D. B. Chem. Soc. Rev. 2002, 31,
342–356.
9. Hartley, D.; Middlemiss, D. J. J. Med. Chem. 1971, 14, 895–
896.
10. Penn, R. B.; Frielle, T. Clin. Rev. Allerg. Immunol. 1996, 14,
37–45.
4.2.7. X-ray powder diffractometry (XRPD). The diffrac-
tion patterns were collected using a Bruker D8-Advance
powder diffractometer, in h–h geometry, using Cu Ka radi-
ation and working at 40 kV and 30 mA. A Sol-Xꢂ solid-
state Si(Li) detector was used. C/Ni Goebel mirrors in
the incident beam were used as a monochromator;
1.0 mm divergence and scatter slits and a 0.1 mm receiving
slit were used, taking care to avoid introducing preferred
orientation of the crystallites.
11. Slattery, D.; Wong, S. W.; Colin, A. A. Pediatric Pulmonol.
2002, 33, 151–157.
12. Agranat, I.; Caner, H.; Caldwell, J. Nat. Rev. Drug Disc.
2002, 1, 753–768.
13. http://www.fda.gov/cder/foi/appletter/1999/20837ltr.pdf.
14. Palacio, M. A.; Cuffini, S.; Badini, R.; Karlsson, A.; Palacios,
S. M. J. Pharm. Biomed. Anal. 2007, 43, 1531–1534.
15. Leger, P. M.; Goursolle, M.; Gradet, M. N. Acta Crystallogr.,
Sect. B 1978, 34, 1203–1208.
16. Jashnani, R. N.; Dalby, R.; Byron, P. J. Pharm. Sci. 1993, 82,
613–616.
4.2.8. Thermal analysis. DSC thermograms were recorded
with a DSC 2920 modulated Differential Scanning Calo-
rimeter (TA Instruments, New Castle, Delaware). Samples
weighing 5–8 mg (Precisa 262SMA-FR Balance) were
17. Ferrayoli, C. G.; Palacio, M. A.; Bresina, M. F.; Palacios, S.
M. Enantiomer 2000, 5, 289–291.
18. Halabi, A.; Palacio, M. A.; Ferrayoli, C. G.; Palacios, S. M.
J. Pharm. Biomed. Anal. 2004, 34, 45–51.